World Development Vol. 31, No. 8, pp.

1343–1366, 2003
Ó 2003 Elsevier Ltd. All rights reserved
Printed in Great Britain
www.elsevier.com/locate/worlddev 0305-750X/03/$ - see front matter
doi:10.1016/S0305-750X(03)00088-3

Transition in Sub-Saharan Africa:

Agriculture, Urbanization and Income Growth
MARY TIFFEN *
Drylands Research, Crewkerne, Somerset, UK
Summary. — Econometric analysis of some 40 years of data has provided mixed results, because of
the defects of the data, and because there are some relatively sudden structural economic shifts. An
important shift is when agricultural labor ceases to grow, now happening in sub-Saharan Africa
(SSA). A model of the interrelationship over time of the rural, mainly agricultural sector, and the
urban, mainly manufacturing and service sector, is proposed. Each provides a market to the other.
Growth in both requires investment, but of distinctly different types. Their interaction results in an
S-shaped curve. Many SSA countries are in the acceleration phase, and its agriculture, particularly
in semi-arid areas, is increasingly oriented to the growing home market. Case studies show farmers
have invested and adopted new technologies but the transition to an urbanized economy has been
hindered by poor policies. The current need is for appropriate investments and policies to develop
the productivity of the urban sectors, so that they can continue to stimulate agriculture, and
provide jobs for those who are leaving farming.
Ó 2003 Elsevier Ltd. All rights reserved.

Key words — transitional economies, development theory, agriculture, industrial policy, statistics,
sub-Saharan Africa

1. TRANSITIONS AND GROWTH investment rate to over 10% of national in-

MODELS
Development theory has been preoccupied
with providing improved living standards and
quality of life either to the great majority, or
more recently, to the poorest of the poor. This
normally requires economic growth greater
than population growth, both to improve
incomes directly, and to increase the taxable
capacity underpinning good government ser-
vices. 1
It is clear that the high-income countries are
urbanized and have large industrial and service
sectors. Agriculture provides only 2% of their
GDP, compared with 10% in middle-income
and 41% in low-income countries (World Bank,
2000b). Most low-income countries have a
higher proportion of their labor force in agri-
culture than its proportion of output value.
Many development economists in the 1950s
and 1960s therefore urged the need to shift
labor into higher productivity sectors, by a
structural transformation leading to industri-
alization of their economies. Rostow (1960)
propounded a rapid take-off stage, for which a
necessary condition was a rise in the productive
1343
come. It was generally believed in the 1960s
that such investment must be made by gov-
ernment, due to lack of private capacity. In
consequence, high taxes on agriculture through
marketing boards and export duties had gen-
eral approval. Helleiner (1966) wrote in relation
to Nigeria:
The disposition of Marketing Board surpluses may
not have been perfect, but the rates of return from
their investments in research, roads, agricultural
schemes, universities, modern manufacturing plants
and so forth are unlikely to have been any lower than
* This paper owes much to the collaboration I have
enjoyed over many years with Michael Mortimore, and
draws upon studies comparing four semi-arid districts in
sub-Saharan Africa, which he led for Drylands Re-
search. These were funded by the Department for In-
ternational Development, UK, which, however, is not
responsible for the views expressed. My intellectual
debts to others too numerous to mention will be par-
tially apparent in the bibliography. They include an
anonymous reviewer who made helpful suggestions.
Final revision accepted: 19 March 2003.
1344 WORLD DEVELOPMENT
those on housing, sewing machines, land clearing, and
the other small-scale outlets for peasant funds dis-
cussed above, let alone so much lower as to offset
the difference between savings rates (p. 184).
The failure of policies aimed at inducing rapid
industrialization discredited belief in a take-off
and stages of growth. 2 The importance of the
agricultural sector as a supplier of raw materi-
als, food and labor, and as the home market
for local industrial output, noted by Johnston
and Mellor (1961), was increasingly recognized.
The low productivity of many government
investments and the outright failure of some
became better known. Structural adjustment
implied ‘‘a shift away from inward-oriented
import-substituting development strategies to
more outward-oriented ones’’ (Alexandratos,
1995), and debt-ridden countries have been
urged to increase and diversify agricultural ex-
ports. 3
If GDP rises smoothly as a result of long-
term effects of capital formation, labor force
expansion and technological change, it can be
modeled by econometric analysis, using the 30–
40 years of statistical data which are now
available for a large number of countries.
Kenny and Williams (2001) have shown
the limited explicatory power of econometric
models, and their often contradictory results.
They note several causes, including, impor-
tantly, the assumption that the process of eco-
nomic growth is the same, not only in all
countries, but in all periods of time. They
suggest therefore, ‘‘that mathematical modeling
techniques have invaded territory to which they
are ill-suited’’ (p. 14).
This paper uses a few generally agreed
principles of economic growth to construct a
nonmathematical model that reflects histori-
cal experience. The model conforms to the
broad thrust of available statistical data, but
shows that economies are fundamentally
different at different points in time, particu-
larly in the relationship of their agricultural
and nonagricultural sectors. Policies therefore
need to be varied according to the stage at
which the country finds itself, rather than
applied universally. The development of the
agricultural sector is initially extremely im-
portant. At a later stage, however, increases
in the productivity of towns are required, to
improve urban incomes, to provide alterna-
tive occupations to the rural poor, and to
stimulate agricultural investment through
growing demand.
Tomich, Kilby, and Johnston (1995) also
returned to a stages approach in their book,
Transforming Agrarian Economies. The model
differs from them in emphasizing the change in
markets rather than in labor disposition, and
the acceleration of change. They identified
countries with abundant rural labor with 50%
or more of the labor force in agriculture, but
thought that
. . .it will be decades before they reach the structural
transformation turning point, when the absolute size
of the agricultural work force begins to decline. Until
then, poverty can be alleviated only if productivity
and employment in the rural economy are increased
(pp. 9–10).
In fact, a number of countries even in sub-
Saharan Africa (SSA) are at or near this point.
Calculating from (World Bank, 2002), urban
population in SSA increased at an average of
5% p.a. and rural at 2% p.a. 1968–2000, and
averaged 34% of the total in 2000. It is now
heading for 50% in some (Figure 1). Rural
population growth had dropped to 0.4% p.a. or
less in eight out of the 20 largest countries in
1998–2000 (Figure 2), and not all rural is ag-
ricultural. While policy debates have focused
on the pros and cons of exports, the swiftly
growing internal market has become much
more important to farmers, and is a main
generator of change in farming systems, par-
ticularly, but not only, in the semi-arid areas.
The urban market is attracting not only their
products, but also their labor.
The model is delineated and explained in
Section 2. Section 3 illustrates the dubiety of
much of the population and national income
data relating to SSA, which lies behind allega-
tions of low or negative growth rates in income
per capita, and falling agricultural output per
capita (World Bank, 2000a). Wiggins (1995,
2000) has already discussed the conflict between
these statistics and the evidence of change and
growth gathered from a limited number of vil-
lage case studies. As these can be unrepresen-
tative, Section 4 illustrates the rapidity of
change in four semi-arid African districts,
1960–2000. Some 30% of the SSA population
live in semi-arid areas (Jahnke, 1982), which
provides a difficult environment for agriculture.
If there are achievements in these areas it is
likely that this is also the case in the better-
endowed areas. Section 5 considers the policies
and government services likely to be most im-
portant in the near future.
 TRANSITION IN SUB-SAHARAN AFRICA 1345

0.8

0.6

0.4

0.2

0
Tanzania
Ethiopia

Uganda

Madagascar

Mozambique

Niger

Angola

Zimbabwe
Nigeria

Kenya

Sudan

Senegal
Ghana

Cameroon
South Africa

Cote d'Ivoire

Mali
Malawi
Congo, Dem. Rep.

Burkina Faso
-0.2

Annual growth %, 1998-2000

Figure 1. Annual percentage growth in rural population, 1998–2000, in sub-Saharan African countries currently having
a population over 10 million, in 1968 and 1998. Source: Calculated from World Bank (2002). Countries in this and
subsequent figures are arranged in order of population size.

60

50

40

30

20

10

0
Uganda
Tanzania

Angola

Senegal
Sudan
Ethiopia

Ghana

Cameroon
Mozambique

Mali
South Africa

Kenya

Madagascar

Cote d'Ivoire

Zimbabwe
Nigeria

Malawi

Niger
Congo, Dem. Rep.

Burkina Faso

1968 1998

Figure 2. Urban population, largest SSA countries, as a percentage of total population. Source: Calculated from the
World Bank (2002).

2. THE MODEL ––The division of labor by specialization

improves productivity and leads to techno-
(a) Basis logical improvements developed out of the
skills and experience of the specialists.
The model utilizes some of the most basic Adam Smith gave his famous illustration
and durable concepts in economics: of pin manufacture. Specialization requires
1346
concentrations of population––in villages
and to an increasing extent as specialization
proceeds, in towns and cities, to facilitate the
exchange of services, products and informa-
tion.
––The division of labor is limited by the ex-
tent of the market. The extent of the market
is determined in great part by the costs of
transport in relation to the value of the
product, and also by the numbers and in-
comes of the potential purchasers. 4
––Improvements in output per labor day or
per hectare require the investment of either
or both capital and labor for a delayed re-
ward, but there are diminishing returns to
additional units of the same input unless
there is also a change in technology or in
the nature of the output. Hence sustained
growth depends on the combination of in-
vestment with new or modified and adapted
technologies (Romer, 1989). This applies
both in agriculture and industry (Anderson,
1990; Tiffen & Mortimore, 1994). The devel-
opment of technology is facilitated endoge-
nously by specialization, but is also helped
by the openness of society to information
from external sources, whether brought in
by the written word and other media,
through the mechanisms of external trade,
100
A
90
80
Slow
70
% agric labour
60 Agricultural labor:
market for urban
50
goods and services
40
30
20
10
0
1
10
19
28
Figure 3. Agricultural, manufacture and service sector labor over time. Note: Arrows show positions of low income and
lower middle income countries in 1960 and 1980, as given in World Bank (1983).
WORLD DEVELOPMENT
or through the concentrations of people typ-
ical of a market place.
––Investment, while a necessary condition
for improved output and incomes, is not a
guarantee of either. A given amount of in-
vestment does not and cannot produce a
given amount of growth. Investment incurs
risk, and is more likely to be successful with
good information about the nature of the
risks and of the market, and if the investor
has been able to develop his/her skills and
judgement. The more complex and larger
the investment, the larger is the range of spe-
cialisms needed to provide information, and
the greater the need for qualitative and
quantitative skills in managing and assessing
it.
(b) Sectoral change over time
There is a symbiotic relationship between the
agricultural and nonagricultural sectors of an
economy, which changes over time, as illus-
trated in Figure 3. In phase A, almost all labor
is absorbed by producing food for its own
household, at low levels of output per labor
unit. In scattered communities in a period when
transport was slow, expensive and difficult,
labor has to be spread inefficiently to cope with
B
C
Fast
Manufacture and service
labor - urban market for ag
produce and ag labor
82
37
46
55
64
73
tim e
 TRANSITION IN SUB-SAHARAN AFRICA
a variety of needs. As Smith (1776, p. 338) re-
marked:
Without the assistance of some artificers. . .the cultiva-
tion of land cannot be carried on but with great in-
conveniency and continual interruption. Smiths,
carpenters, wheelwrights and ploughwrights, masons
and bricklayers, tanners, shoemakers and tailors are
people whose service the farmer has frequent occasion
for.
Farmers are unable to sell any surplus they
produce, if all near them are similarly engaged,
and they have no access to other centers of
demand. 5 Typically, occasional surpluses are
spent in feasting and drinking rather than in-
vesting to produce more. With low population
density, extensive agriculture, using long fallows
to restore fertility, is appropriate (Boserup,
1965).
Moving out of this situation is slow and
difficult. Figure 3 shows that agricultural and
nonagricultural labor form each otherÕs mar-
ket. The respective size of the market, shown by
the vertical columns, limits the development of
the other sector. Typically the first nonagri-
cultural sectors to develop are those of ad-
ministration and defense, which often cull any
small agricultural surplus that exists without
contributing much to the reduction of ‘‘incon-
veniency’’ and the interruption of agricultural
tasks by other necessities.
The units of time in Figure 3 are not stan-
dardized. Phase A typically lasted many cen-
turies while the transition phase B may be
accomplished in 50–500 years. Its length, often
difficult to document in the absence of histori-
cal data, depends in part on the technologies
available in an external contact economy
(Gerschenkron, 1962). In the early slow part of
phase B it is usually assisted by access to ex-
ternal capital and markets to expand the non-
agricultural sector and transport facilities. In
Britain, generally recognized as the home of the
industrial revolution, it began with raw wool
exports to only slightly more advanced econo-
mies in Europe in the medieval period, and
started gathering speed circa 1760–1800. Later
developing countries have generally made the
transition more quickly than the United King-
dom. The most recent examples are in eastern
Asia. 6 Change accelerates as productive tech-
nologies and technologies for the exchange of
goods and information improve. Concentra-
tions of people facilitate the exchange and de-
velopment of ideas (Simon, 1977). Countries
1347
that cut themselves off from the exchange of
goods, ideas and capital, as did China for many
centuries, (Landes, 1998), and Albania more
recently, move more slowly down the curve.
The contact economy also provides a market
for farmers hitherto dependent on an extremely
small nonagricultural sector. The enlarged
market provides a ‘‘vent’’ for farmers, and
stimulates efforts to provide a surplus above
family needs in return for desirable consumer
goods and inputs, many of which are initially
imported. In countries which initially have
some underused labor and land the export crop
is additional to existing production (Myint,
1967). Hopkins (1973) illustrated this with
special reference to West Africa. In countries
with large numbers of peasant farmers, their
additional purchasing power stimulates the
consumer goods and service sectors. This effect
is not so powerful in countries where a few
people control most of the land and hold down
agricultural wages, as in parts of Latin Amer-
ica, as Tomich et al. (1995) point out. Fortu-
nately, this does not apply in most of SSA,
where small-scale farming predominates.
The external economy may also provide in-
vestment capital that begins to expand the local
urban sector, with port and transport facilities,
etc. Smith thought the rapid progress toward
wealth of ‘‘our American colonies’’ in the 17th
and early 18th century was due to the fact that
their citizens were able to concentrate their
scarce capital resources on the development of
agriculture (which he saw as its most produc-
tive and secure use) because British merchants
were supplying the capital for the much more
risky export trade––ships and even warehouses
(Smith, 1776, p. 329). The steam ship and
railway expanded the reach of export markets
in Asia in the 19th century, but, aside from
ports, major transport and urban infrastructure
investments for SSA only occurred in the first
half of the 20th century.
Under the stimulus of export markets for a
product which many can supply, agriculture
begins to become more productive. Initially,
there is vacant land farmers can develop for
cropping. They are helped by better implements
usually brought in by trade––axes, metal hoes,
animal-drawn ploughs, etc. These are often
locally manufactured once the market is es-
tablished, though they will only be selectively
adopted, according to their suitability for the
land-labor ratios and other local conditions
(Pingali, Bigot, & Binswanger, 1987). 7 Farm-
ing can therefore begin not only to supply food,
1348 WORLD DEVELOPMENT
but also to release labor to a growing urban
sector, which is beginning to produce and
supply consumer goods, services and some of
the inputs for agriculture, reducing the incon-
veniency of self-manufacture, and freeing up
labor for more intensive farming. The urban
sector can begin to grow because farmers are
now becoming a local, effective, cash-earning
market. The economy is embarking on phase B,
in Figure 3.
In turn, a larger, more productive urban
sector enlarges the market for farmers and
stimulates them to invest in improvements.
Without investments to improve productivity a
falling proportion in farming would be unable
to supply the needs of a greater proportion in
the nonagricultural sectors. The internal mar-
ket becomes very large by phase C. At this
stage the manufacturing and service sectors are
no longer limited to the agricultural sector for
their internal market, since they are also selling
to other specialists in the nonfarm sectors. The
size of the internal market for farm output
depends on the incomes (and therefore pro-
ductivity) of urban people, as well as actual
numbers in towns, since BennetÕs Law predicts
that as incomes rise, the share of starchy staples
in food declines, and diet shifts toward live-
stock products, fruit and vegetables (Tomich
et al., 1995, p. 163). It is also governed how-
ever, by the costs and speed of transport facil-
ities for bulky and/or perishable commodities. 8
Rural should not be equated with agricul-
tural. While agricultural activities take place
almost entirely in the rural areas, rural areas
will also develop services and small-scale man-
ufacturing activities to meet local needs. Large-
scale manufacturing industries will be located
in urban areas at nodal points on the transport
network, to facilitate distribution to a wide
market. 9
As the manufacturing and service sector in-
creases its productivity and income, it attracts
labor out of agriculture, leading to the faster
stages of phase B. At some point, even in
populations that are still growing, the actual
numbers of people performing agricultural
tasks begins to drop, as well as their share in
the work force. There is a growing need for
additional capital to substitute for labor. As an
example, farmers on a new irrigation scheme in
Muda, Malaysia, found themselves obliged,
during 1974–79, to begin hiring combine har-
vesters and other machinery to replace the
labor of sons and daughters who had moved
into a growing industrial sector. 10 Some
farmers are forced out of farming by their in-
ability to finance new capital requirements, or
their heirs choose to sell up to transfer into
activities perceived as more profitable. A ten-
dency toward larger, capital-intensive farms
develops as phase C approaches. This paper is
not examining phase C, but concentrates on
phase B, as it is likely to develop in SSA.
(c) The importance of investment and the
differing nature of investments required to
improve agricultural and nonagricultural
productivity
While in historical experience the growth of
the manufacturing and service sector is gener-
ally associated with higher incomes per capita,
this desirable outcome has depended on effec-
tive investments in all sectors, and, therefore,
on policies that have provided incentives and
security for investment in and maintenance of
new assets, and the availability of information
and education to increase ability to select ef-
fective investments.
Characteristically, but not exclusively, raising
the productivity of the small-scale farming
sector requires many repeated small private
investments (developing new land, acquiring a
new tool or other input, increasing the value of
livestock held, planting and nurturing tree
seedlings, etc.). These are barely perceptive to
outsiders, though they can have a significant
impact on production (Tomich et al., 1995, p.
21). In some cases the investment can be of
work for a delayed reward, rather than actual
cash. The cumulative size of these incremen-
tal and intermittent investments depends on
whether there is an enabling policy environ-
ment, which preserves private incentives to in-
vest, spreads information about opportunities
and risks, and improves access to markets. Ef-
fective private investment in agriculture is most
assisted by appropriate public investments
outside agriculture, for example, in communi-
cations infrastructure 11 to assist marketing
and the gathering of new information. State
investments in primary education can assist in
the uptake and assessment of information.
Although agricultural investments are typi-
cally small and incremental, there are excep-
tions. Government investment in large irrigation
facilities has been crucial in Asia, though often
beset by difficulties in management and main-
tenance. The water resources of SSA make it
necessarily more reliant on small-scale irrigation
to which farmers can themselves contribute re-
 TRANSITION IN SUB-SAHARAN AFRICA
sources and management inputs. State invest-
ment in research has also been important and
can speed up the development of improved crop
varieties, etc.
By contrast, large increases in the produc-
tivity of the industrial and service sector depend
characteristically, but not exclusively, on large-
scale, lumpy, investments in electricity, water,
communications, etc., which are typically sup-
plied by state, municipal authorities or by
shareholders in large private firms (often for-
eign in the initial stages). These are being
termed collective investments, in contrast to the
family-scale investments in farming. Large
collective investments, like family investments,
need good information and risk assessment to
succeed but, as the demands are more complex,
this requires public and private investments in
higher education for management as well as in
primary education for the workforce.
(d) Goodness of fit with statistical time series
It has already been said that the curve in
Figure 3 is not based on consistent time units.
It is presented to illustrate a pattern, and the
steepness of the slope in phase B will vary ac-
cording to the circumstances and policies of
particular countries. The point at which it levels
out will depend in part on the size and natural
assets of different countries, and in part on their
policies. But, it will always take the form of a
backward sloping S, owing to mutually rein-
forcing and accelerating impacts of one sector
upon the other. These impacts are related to the
efficiency and income effect of the investments
made in each sector. They will be less where the
major element in the service sector is an un-
productive civil service, or where the state has
invested in manufacturing projects that fail to
give an economic return. The World Bank es-
timates that the productivity of investment in
SSA needs to double (World Bank, 2000a).
Nevertheless, the model as depicted fits rea-
sonably well with data on percentage of the
labor force engaged in agriculture for 1960 and
1980. The change in their position is shown by
the arrows on the curve. Low-income countries,
less than 18% urbanized in 1960, moved rela-
tively slowly in 20 years from 77% engaged in
agriculture to 70% in 1980. Lower middle-
income countries, less than 38% urbanized in
1960, moved more rapidly in the same 20 years
from 71% in agriculture to 55%. The figures are
not exactly comparable, as some countries have
moved from one category to the other in the
1349
same 20 years. They tend however, to confirm
the thesis that urbanization and, by implica-
tion, the development of the manufacturing
and service sector, accelerate, giving rise to the
S-shaped curve depicted in Figure 3. They also
confirm that this process is associated with in-
creasing wealth per capita. The agricultural
labor force figures are somewhat suspect, since
they tend to be worked out as a residual after
the known workforce in formal sectors have
been deducted (Tomich et al., 1995). They in-
clude many in the nonformal nonagricultural
sector. Hence, the data on urbanization provide
a better guide to the scale of change, though
there are problems with varying definitions of
urban. Figure 2 shows that since 1960 most
SSA countries have become over 30% urban
and five are over 40%. This implies agricultural
labor occupies no more than 60–70% of the
labor force, placing them clearly in the transi-
tion phase B of Figure 3.
3. THE WEAKNESSES OF THE
STATISTICAL BASE
National statistical data have many weak-
nesses in SSA (and probably elsewhere).
(a) Population and labor data
While the World Bank population and ur-
banization figures can be accepted as giving
orders of magnitude, they are not safe as ac-
curate inputs into mathematical comparisons
between countries. 12 Censuses have been in-
frequent in many SSA countries. The results are
late in entering national and international sta-
tistics. The on-going population transition in
some countries, and the AIDS epidemic in
others, have probably lowered annual growth
rates below estimated rates. The most glaring
errors are in Nigeria, where the census of 1991
counted 88.99 million. However, the World
Bank (2002) was still giving 98.98 million as the
1991 figure, and adding 2.8% annually till 1997,
reducing this by stages to 2.4% in 2000. This
produced an estimated population of 127 mil-
lion in 2000, whereas a continuous growth rate
of 2.4% from the census figure results in 110
million. 13 The discrepancy in the divisor affects
all Nigerian per capita data, and even affects
group data for SSA, as Nigeria holds 20% of
the SSA population.
Definitions of urban vary by country and
by census date. Urban in Nigeria in 1963 was
1350 WORLD DEVELOPMENT

defined as agglomerations over 5,000, and over local retailing, transport, and manufactures
20,000 in 1991. In Niger, in 1988, ‘‘urban’’ was rises when farming areas generate higher out-
places of more than 2,500 with a government put and higher incomes. In five villages in Java
office (Republic of Niger, 1992). In Senegal, visited in 1974 and revisited in 1987, it was
urban in 1988 meant having a municipal com- observed that farm incomes had improved by
mune. The second largest city, Touba, a reli- possibly 30–50% in real terms, and there was
gious rather than administrative headquarters, better clothing, wider possession of consumer
was administratively rural, though it counted goods, and better access to education and
183,000 inhabitants in 1988, (Barry, Ndiaye, transport (Prabowo & McConnell, 1993). This
Ndiaye, & Tiffen, 2000). By 1999 it was re- obviously meant more employment of rural
ported as 500,000 (Coulon, 1999)––not insig- teachers, drivers, shopkeepers, etc. and there
nificant in a country with a total population of were specialized industries in some villages. On
10,000,000. the Muda irrigation scheme in Malaysia every
Employment by sector is generally acknowl- additional $1 earned from improved rice output
edged to be the weakest part of most census generated another $0.74 of activities and earn-
data. 14 The percentage living in rural areas can ings off-scheme (Bell, Hazell, & Slade, 1982).
be taken as the upper limit for agricultural la- These intersectoral linkages are also noted in
bor, since the number of rural people whose SSA (Barrett, Reardon, & Webb, 2001) but
main activity is outside farming is almost cer- these often ‘‘informal’’ activities may not be
tainly greater than the number of urban people caught in official statistics.
whose main activity is farming. Many studies
have found considerable rural nonfarm income. (b) Agricultural output, food imports and
In a review of 25 case studies in SSA, at times contribution to the GDP
varying from 1974 to 1990–91, Reardon (1997)
found its average share of household income to Agricultural output is notoriously difficult to
be 45%. Bryceson (2002) found higher levels by estimate. The amounts consumed on farm or
the later 1990s in six SSA countries, with 60– traded domestically are likely to be underesti-
80% of income from nonagricultural sources. If mated. Figure 4 shows the value of agricultural
carried out only in the off-season these activities exports as a percentage of agricultural GDP for
are not competitive with farm labor, but some 1995. It is immediately apparent that Zimba-
household members give preference to nonfarm bwe and Malawi were not valuing their locally
labor even in the farming season. Demand for consumed crops and livestock. 15 A survey in a

100
90
80
70
60
50
40
30
20
10
0
Uganda
Tanzania

Angola

Niger
Nigeria

Ghana

Cameroon
Congo, Dem. Rep.

South Africa

Kenya

Madagascar

Cote d'Ivoire

Zimbabwe

Mali
Burkina Faso

Malawi
Mozambique

Agr. exports as % agr. GDP

Figure 4. Agricultural exports as a percentage of agricultural GDP in largest SSA countries, in 1995. Source: Con-
structed from World Bank (2000) (data not available in 2002 version).
 TRANSITION IN SUB-SAHARAN AFRICA
tobacco growing area of Malawi found that
just over 40% of farmersÕ income came from
crop sales, of which only 15% in 1990 and 9% in
2000 came from the export crop, tobacco. The
field pea contribution was 10% in 1990 and 23%
in 2000 but until 1998 this crop was not offi-
cially recorded (Orr & Mwale, 2001).
Despite the growing urban population, who,
being slightly better off, consume more high-
value foods and beverages than rural people,
food imports have generally remained in a
range of $3–$10 per head (Figure 4) since
1980. 16 Food import data in countries with
limited import routes are likely to be more re-
liable than food production data, but not all
countries measure imports. Calculations from
(World Bank, 2002) show that among the nine
large countries supplying this data, in eight,
food imports per capita have been static or
falling since the mid 1980s, Kenya being a
possible exception. In most they have remained
below $10 per capita, in constant 1995 US$.
Senegal, where policy favored rice imports to
increase specialization on groundnuts, is ex-
ceptional with a level generally above $25 per
capita, though it seems to have fallen since the
devaluation of the FCFA in 1994. Nigeria, up
at nearly $35 in 1981–82, fell back to $4–8 after
successive devaluations. Some temporary rises
due to droughts can be seen, particularly
marked for Zimbabwe, 1989–91. The implica-
tion is that in most countries, farmers have kept
up with the level of demand, and changes in its
nature, with variations in part due to policies,
and occasionally to drought. This is incom-
patible with the calculated volume of food
output by major crop given in Tables 8–6
(World Bank, 2002), (which quotes FAO data).
This shows annual percentage growth in almost
all cases of less than 1%, which is certainly less
than population increase. Static per capita
production implies an annual growth in food
crops equal to population growth, unless im-
ports are rising. 17
Farmers might meet internal demand with-
out increasing productivity if they transfer re-
sources from export crops to food crops. But,
volumes for three main exports, cocoa, cotton
and coffee, 1968–97 were maintained (World
Bank, 2002). Groundnut exports fell sub-
stantially during 1968–78, stabilized 1978–88
and fell again somewhat during 1988–98. Our
studies show more of the crop going to meet
local demand, but falls in export prices and
rosette disease in Nigeria in the 1970s are other
factors.
1351
If we ignore the countries with obvious errors
in Figure 4, most countries are not exporting
more than 20% of agricultural production
value, so 80% is locally consumed. Crops in-
ternally consumed are now much more impor-
tant than crops exported, as we should expect
to happen with countries in the transition phase
B depicted in Figure 3.
4. FARMERS’ INVESTMENTS IN
RESPONSES TO CHANGING MARKETS
IN FOUR SEMI-ARID DISTRICTS OF SSA
Given the deficiencies in national data, the
processes of change are best understood and
tested against the model in Figure 3 at a district
level. At this level sample data from villages can
be related to district statistics and, therefore, to
the national statistics built up from them. This
has been done for four semi-arid areas, 1960–
2000. These are Makueni District, Kenya,
Diourbel Region, Senegal, Maradi Depart-
ment, Niger and the Kano hinterland, Nigeria.
Farmers in these areas are restricted in the
crops they can grow, unless they have access to
pockets of irrigable or water-retentive land.
The main crops are cereals and pulses, low in
value in relation to bulk, but capable also of
providing fodder to livestock. In the past the
main export crop was groundnuts, and in
Makueni, cotton. Three districts have one short
farming season lasting three to five months, but
Makueni has two wet seasons of about three
months each. Rainfall is very variable from
year to year, giving a risk of crop failure.
Nevertheless semi-arid areas contain large
populations, with particularly high densities in
the hinterlands of Kano in northern Nigeria
and Dakar in Senegal.
In terms of the model, we are particularly
interested to see whether there has been esca-
lating growth in urban populations. If so, we
need to see if farmers have been making the
appropriate private investments to respond to a
growing urban market for both their products
and their labor, or if they have been held back
by deficiencies in policy, or in the government
supply of some of the services identified as
particularly necessary. Changes in land markets
should also occur as land becomes more scarce,
and in labor disposition if urban occupations
become more attractive. Second, we need to see
if the collective investments and policies re-
quired to make the growing nonfarm sector
more productive are in place, so that it can
1352 WORLD DEVELOPMENT
both provide jobs for people leaving the farms,
and a growing market for higher value farm
produce such as livestock, fruit and vegetables.
The studies were carried out by teams of
scientists from the countries concerned, profil-
ing particular aspects of change, 1960–2000,
using methods and data appropriate to the
subject matter. This included literature reviews;
collection, discussion and analysis of district
data with the appropriate government office;
air photo and remote sensing interpretation;
and collective interviews and sample surveys of
10–12 randomly selected farmers in four rep-
resentative villages. 18 The country team lead-
ers were Francis Gichuki (Kenya), Abdou Fall
and Adama Faye (Senegal) and Yamba Bou-
bacar (Niger), who collaborated with Drylands
Research in developing from the profiles three
national syntheses, using the findings to trace
and understand socioeconomic and environ-
mental change in its various interactions over
time. The Nigerian study was limited, for fi-
nancial reasons, to five profiles, which included
an in-depth study of the marketing of food
crops and livestock in the Kano area and a
review of policy. Findings were checked by one
or more workshops in which District officials
and farmer representatives took part. DFID
also provided funds for endorsement exercises
in the local constituencies in 2001–02.
(a) Changes in rural and urban population and in
urban demand
Rapid growth of the local towns occurred in
three of the four study areas, with urban pop-
ulation being now 30% or over in two cases
(Table 1).
The increase in the urban market over time is
particularly visible in northern Nigeria and
Senegal, but the outcome in terms of effective
demand for local farm produce differed. Table
1 shows the increase in local urban grain re-
quirements estimated at 200 kg/head. During
1952–91 the urban population of the two states
constituted from the colonial Kano Province
increased at about 8% per year, while the rural
increase was nearer 1.6%. In 1952 Kano mu-
nicipality was the only town with more than
20,000 inhabitants; by 1991 it had been joined
by 11 others (Tiffen, 2001). To cater solely for
the local towns, each rural family needed to
supply 10 times as much grain to urban mar-
kets on average in 1991 compared with 1952. In
addition, they were supplying grains and pulses
(especially maize and cowpeas) to the even
more urbanized southern Nigeria (Ariyo, Voh,
& Ahmed, 2001). If Nigerian farmers as a
whole had failed to invest, food imports would
have increased dramatically, but this has not
been the case. In the late 1960s NigeriaÕs food
imports were $2–$3 per capita (World Bank,
2000a, 2000b, 2000c). Figure 5 has shown that
although these billowed during the petroleum
boom and while the currency was overvalued,
in the 1990s they were running at only $5–$8
per capita (all in constant 1995$). Given that
the import figure includes all foods and bever-
ages, this does not suggest a large import of
staple grains. The scale of the increased pro-
duction demanded from farmers is understated
in Table 1, as not all families living in towns of
less than 20,000 are primarily engaged in agri-
culture. It also takes no account of industrial
needs. Maize, sorghum and millet are impor-
tant inputs into large commercial brewing,
flour milling and animal foodstuff enterprises
(Swindell, Iliya, & Mamman, 1999).
Crop production data for the Kano area was
only available to us 1982–90 (World Bank,
1995), when yields were strongly linked with
rainfall (Tiffen & Mortimore, 2002). But, for
Maradi, Niger, there is a series of District fig-
ures running from 1964 to 1998. These show
the expected variation in production per capita
(taking the total population) from year to year,
under the influence of rainfall variation. They
also show that, after recovery from the
droughts in the early 1970s, farmers upped their
production to average around 300 kg/capita,
(Mortimore, Tiffen, Boubacar, & Nelson, 2001)
despite a growing rural population, and despite
the virtual cessation of supportive projects after
1985, under structural adjustment (Hamadou,
2000a; Figures 4 and 5). This must mean that
they saw an incentive to increase their sales.
The net sellers of grain to the towns are not
evenly distributed among the rural population.
In a large CARE sample in Maradi in 1996 the
families defined as the most vulnerable (56%)
produced 125 kg cereals/capita, against a
moderately poor group (27%) who produced
311 kg, and a relatively well-off group of 17%
who produced 580 kg/capita (CARE Interna-
tional au Niger & Universite dÕArizona, 1997).
The average was 252 kg/capita, in a somewhat
below average rainfall year. Food selling was
near universal but if one estimates per capita
needs at 200 kg/capita/annum, it is apparent
that quite large numbers of rural families are
net buyers of cereals, and that a minority are
net sellers. 19 Maradi town has grown, but
 Table 1. Population and urban grain demand per rural person in four semi-arid areas
Total Urbana % Urban Urban growth p.a. Rural density Urban grain need, kg
per rural personb

TRANSITION IN SUB-SAHARAN AFRICA

Kano Province 1952 3,396,350 130,173 3.8 77 8.0
Kano and Jigawa States 1991 8,685,995 2,516,686 29.0 7.9 (Jigawa) 118 81.6
(Kano) 169
Maradi 1960 (estimate) 561,000 – – 13
1977 949,747 44,459 4.7 – 22 9.8
1988 1,389,443 110,739 8.0 8.7 33 17.3
Diourbel 1960 (estimate) 261,000 n.a. n.a. n.a.
1976 423,038 120,249 28.4 70 79.4
1988 620,197 251,799 40.6 6.9 94 136.7

Makueni (dryland only) 1962 170,717 – – – 38

1989 524,025 – – – 102
Source: Calculated from census data: summaries in Gichuki (2000a), Barry et al. (2000) and Tiffen (2001).
a
The 1991 Nigerian urban definition as settlements of over 20,000 is here applied to all data, regardless of country and time differences. Makueni has no towns of this
size.
b
Grain need assumed as 200 kg per head per year. Urban demand at this level has been divided by rural population. To get demand per rural family it is necessary to
multiply this up by family size, which varies by country.

1353
1354 WORLD DEVELOPMENT
Figure 5. Structure of incomes by type of household, Diourbel, 1999–2000. Source: Faye and Fall (2000), Figure 1.
Household types as in Table 2.
important influences on farmers are the large
towns across the border in Katsina State, Ni-
geria. Grains and pulses flow in both directions,
depending on harvest conditions, prices, and
exchange rate differences (Meagher, 1997).
Most Maradi farmers have relatively small
farms. Cereal yields are in the order of 300–500
kg/ha, which gives an idea of the modest
cropped hectares in families producing 125–600
kg/head. The distinguishing characteristic of
large sellers is their good management quality,
which has enabled them to accumulate re-
sources (from livestock sales, crop production
or nonfarm incomes) for investment in inten-
sification. Farmers of one southern Maradi
village said, in discussing the findings of the
research, that farming was only profitable for
those who had the means to invest, and those
who could await a favorable moment for sell-
ing. Commerce, crop farming and livestock
raising could not be disassociated (Boubacar &
Ibrahim, 2002).
In northern Nigeria the most efficient farms
were found to be most numerous in high-den-
sity areas with good market facilities. These
had smaller than average farms with 4.4 ha as
against the average in the sample of 5.8 ha
(Okike et al., 2001). There are larger farms in
the vicinity of some big cities, encouraged by
the state, where entrepreneurs, variously from
trade, the old aristocracy, modern business or
the civil service and army, bought land in the
1980s and invested heavily in mechanized farms
using hired labor, in response to a perceived
growing and profitable demand for millet and
other staples. In the 1990s their average size in
the vicinity of Sokoto had been reduced to 20–
50 ha because of rising costs of fuel, shortage of
spare parts, and high wage bills (Swindell et al.,
1999), a trend also noted by Balcet (1997). A
sudden burgeoning of large-scale farms was
also seen round Kano but they were estimated
to account for only 5% of total agricultural
production. Mustapha and Meagher (2000)
noted that many of these ‘‘modern’’ enterprises
foundered when state support was reduced in
the aftermath of falling oil prices. A particular
feature in northern Nigeria has been the bur-
geoning of fadama farming (on low-lying or
riverine land), utilizing small pumps to raise
groundwater, to irrigate wheat and vegetables.
This demands large inputs into small areas
(generally less than one ha) but leads to very
profitable sales to the towns (Swindell et al.,
1999; World Bank, 1995).
Local urban growth was also very substantial
in Diourbel, Senegal, due largely to the new city
of Touba as well as growth in the two of the
three older departmental towns (Barry et al.,
2000). In addition, farmers had easy access to
the expanding capital, Dakar. But, there was
almost no urban market for the grain they
could produce, millet, due to the long-estab-
lished preference for rice. While they main-
tained output of millet per agricultural worker
1960–92 at around 500 kg for family use
(weather permitting), output per total popula-
tion fell to around 100 kg (Faye, Fall, Morti-
more, Tiffen, & Nelson, 2001, Figure 14). The
market for groundnuts as a local cooking oil
and ingredient was distorted by government
efforts to channel all groundnuts to state oil
mills, whose high processing costs mean that
urban consumers preferred cheaper imported
oils (Gaye, 2000). Hence, the main local impact
of urbanization was on demand for meat, par-
 TRANSITION IN SUB-SAHARAN AFRICA
ticularly after the devaluation of the FCFA
made imported meat more expensive. From
erratic figures available from the veterinary
department, small livestock increased from
100–140,000 1966–72 to 200–250,000 in the
1990s. (Cattle numbers were relatively static;
holdings of equines grew somewhat, Faye, Fall,
& Coulibaly, 2000). Sales were not equally
distributed, for it was only the farmers who had
the resources to invest in livestock, concen-
trated among those who did not usually have to
buy in grain, who were able to go in for sub-
stantial livestock fattening (Figure 5). As in
Nigeria, there are some much larger, more
mechanized farms, operated mainly by religious
leaders and businessmen, whom we excluded
from the sample. In the early 1970s these only
produced a small proportion of the groundnut
crop (OÕBrien, 1975), but the current position is
not known.
The exception to local urban growth was
Makueni, in Kenya. This part of the former
Machakos District had experienced an influx of
rural settlers 1950–90 (Tiffen, Mortimore, &
Gichuki, 1994). Neither Wote, which was made
the capital of the newly formed Makueni dis-
trict in 1992, nor any of the settlements along
the Nairobi–Mombasa highway on its southern
edge, had attained a population of 20,000 by
the 1989 census. It may be significant that,
unlike the district towns in the other three
countries, there had been no collective urban
investment to 1998 in electricity, telephones,
etc., 20 so, even on the highway, enterprises
were limited to services to travelers, marketing,
and such crafts and small industries as can
operate competitively without power.
In Kenya as a whole, Figure 2 shows that
urbanization was estimated at 30% by 2000.
Makueni farmers had some access, over poor
roads, to Mombasa and Nairobi, for their
higher value crops such as fruit and vegetables
but had no comparative advantage in produ-
cing maize for these towns (Gichuki, Mbogoh,
Tiffen, & Mortimore, 2000). Mbogoh (2000)
found cereal production per person since the
district was formed in 1992 varied enormously
from year to year, according to district statistics
(which are no more than guestimates). 21 The
average grain and pulse production 1992–98
just about met district needs taking all years
together. In some years however, farmers were
net buyers of cereals and beans (and forced
sellers of livestock) while in other years they
could sell grains and restock on animals. Un-
fortunately, livestock disease could wipe out
1355
their insurance against crop failure (Fall, 2000).
In the worst years, many were dependent on
food aid from relatives or the government.
(b) Farmers as investors
As business managers, farmers have to decide
how to react to changing product and factor
markets, and whether, and in what, to invest.
In northern Nigeria growing food demand
has been met by complementary farm invest-
ments, despite having to cope with frequent
changes and reverses in government policies
which have affected the level and competitive-
ness of food imports, the cost and availability
of farm labor, the subsidization or not of fer-
tilizer and other inputs, etc. (Mustapha &
Meagher, 2000). In response to the labor
shortage induced by the urban-centered infra-
structural petroleum boom in the 1970s, Kano
farmers invested to a very considerable extent
in buying or hiring ox-drawn equipment, pre-
viously rare in the Kano-close settled zone.
Maize, a minor crop in 1979, became a major
crop in suitable areas by 1989, as farmers
adopted a new variety (TZB) that was very
responsive to fertilizer, subsidized at the time
(Smith, Barau, Goldman, & Mareck, 1994).
When the fertilizer price was raised, farmers
reduced its usage, 22 and reverted to older tech-
niques of maintaining fertility, long practiced in
the densely settled area around Kano (Morti-
more, 1993). They selected and crossed new and
old varieties of seed that best matched their
ecological niche and changed economic cir-
cumstances (Mortimore & Adams, 1999), while
also adopting new crops such as soya beans
(Balcet, 1997).
Farmers have also responded to the growing
demand for meat. The most efficient farms,
with the highest gross revenue per ha and the
highest net revenue per farm, were those pur-
chasing in the most crop residues and other
inputs, to maintain more than double the av-
erage number of livestock in addition to crops,
using three times as much manure as the aver-
age and more days of animal traction (usually
hired) (Okike et al., 2001). Areas of common
grazing in the densely settled Kano and Jigawa
states are very limited, but 1984 data show the
high numbers of livestock kept on farms with
some water-retentive land (20–40% of such
households had cattle, and 75–90% had around
15 sheep and goats) (World Bank, 1995, Table
3.8). The livestock are penned to maximize
manure collection and to fatten up quickly. The
1356 WORLD DEVELOPMENT

high inputs of labor this requires for feeding
and watering, as compared with herding
(Mortimore & Adams, 1999) is only worthwhile
if there is an active market for livestock as well
as cereals. There is now an active market for
crop residues, with prices higher for groundnut
and cowpea hay than for grain (Baba & Mag-
aji, 1998). Farmers have recently shown interest
in a new cowpea variety that gives more fodder,
though it requires spraying (Singh, Larbi,
Tabo, & Dixon, 2001).
Farmers in Diourbel, Senegal, also calculate
the returns to different farm enterprises and
technologies and invest accordingly. In the
1960s through to the 1990s, Senegal farmers
weighed up the value of investing in fertilizer in
relation to its price and groundnut prices, and
the merits of heavy versus light ploughs, or in
buying a plough versus a cart for transport
(Faye et al., 2001). In response to the growing
urban market for meat, and the rising price of
mutton, many in this so-called groundnut basin
had by the 1990s decided to go in for animal
fattening. Schoonmaker Freudenberger and
Schoonmaker Freudenberger (1993) reported
how families in a village to the north of
Diourbel said it was better to invest in five
lambs than to seed an area with 500 kg of
groundnuts. Livestock have become a major
element in farm incomes of the better off, as
shown in Figure 5. By 1999 much of their
groundnut production was not sold for export,
but consumed locally as nuts or oil, with the
cake and hay providing valuable fodder sup-
porting their livestock enterprises (Faye, 2002;
Faye & Fall, 2000). Their investments have
followed suit, as shown in Table 2. It is not
surprising that the poorest spend little on farm
inputs, but it is worthy of note that the in-
vestments of the middle and better-off segments
get directed toward livestock rather than crop
enterprises.
The way farmers plan, assess, invest over
time and juggle resources, was particularly il-
lustrated in the Makueni study. Farmers were
asked to identify the three most important past

Table 2. Use of money income by 38 Diourbel farmers (1999–2000)

a
Type of household Type 1 Type 2 Type 3
Household characteristics
Percentage of sample in this type 29.4 41.2 29.4
Estimated income per head in past year (FCFA) 46,456 64,054 91,019
Months own millet and groundnuts met family needs <6 6–11 12+
Percentage of income not farm generated 68 48 18
Percentage of family workers absent in farming season 25.0 16.7 18.8

Use of money income (per cent)

Consumption
Buying millet 16.9 7.5 5.4
Buying rice 15.2 13.1 10.3
Buying other food 22.0 11.7 26.6
Clothing 14.3 19.2 9.0
Buying other products 5.6 2.7 3.2

Farm-related expenditure
Equipment 0.5 0.6 2.1
Buying animals 0.8 12.0 18.9
Buying animal food or medication 1.1 4.4 2.9
Buying crop inputs 1.9 1.9 2.4
Payment for services 1.6 1.8 2.0
Other expenditure
Health and education 3.6 4.1 2.7
Ceremonial expenses 16.6 21.0 14.5
Total 100.0 100.0 100.0
Source: Faye et al. (2001).
a
Percentage not farm-generated calculated from original data not there quoted. The sample was small, being based
on 10 randomly selected farmers in four villages but the results are all in the direction expected.
 TRANSITION IN SUB-SAHARAN AFRICA 1357

Figure 6. Expansion of cultivation in four terroirs, Maradi Department, Niger, 1957–99, in four cultivation frequency
classes (percent of area). Source: Mahamane (2001): Tables 5, 7, 9, 11. Villages arranged from south to north. No data
are available for Jiratawa in 1957.

investments contributing to their welfare. Ter-
racing, planting trees, clearing bush, building a
house, and educating children topped the
charts (Nelson, 2000). Clearing bush and
building a house were preliminary investments.
Terraces, trees and education came in later
stages. On the farms in the sample, the con-
struction of terraces and drains had taken, per
ha, 30–60 man days and Kenya shillings 5,000–
9,000 (Gichuki, 2000b), and had been gradually
extended to all their cropped land. All farmers
had planted fruit trees, ranging from three on
very small farms to over 200 on some larger
ones. This involves considerable investment in
labor for planting pits and watering during es-
tablishment, as well as manure and seedling
purchase. Half the sample had made invest-
ments intended to improve output even during
the difficult 1990s, bunched in years giving a
good harvest. All farmers could readily identify
the next one or two investments they wanted to
make when resources became available. These
were mainly related to equipment (67%), inputs
(17%) and granaries (13%), needed to increase
their output of cereals and pulses (Mbogoh,
2000). A good half of the finance for their in-
vestments came from nonfarm income (Nelson,
2000), which in bad years gets taken up by
consumption needs.
In Maradi, the main investment initially was
in clearing new land for cultivation, within the
village area while available, later further north.
The process is made visible by air photograph
and satellite data (Figure 6). In one of the four
villages, Jiratawa, there was also project in-
vestment in an irrigation scheme, but the others
were unassisted. Owing to the availability till
recently of land that could be fallowed or
cleared, investment in intensification is only just
beginning. Investments in equipment (Table 3)
have accelerated since the devaluation of the
FCFA, which, as in Burkina Faso (Hoffmann,
1998) and Senegal, has resulted in improved
prices for local livestock and grains (Kherallah,
Delgado, Gabre-Madhin, Minot, & Johnson,
2000).
The volume of farm investment in the dis-
tricts over the 40 years could not be calculated,
but it was certainly substantial in Makueni,
Maradi and the Kano area. In Makueni and
Maradi, credit was unimportant, and finance
came from crop and livestock sales, family
transfers, and use of nonfarm income. We have
no data on the role of credit in the Kano area,
but some was available. Credit was part of the
state-imposed system in Diourbel, c1965–80,
when all producers had to belong to a co-
operative, which provided inputs on credit, and
sold output at controlled, low, prices. In bad
years, farmers were obliged to default, eventu-
ally bringing about the bankruptcy of the re-
sponsible state organization (which was also
1358 WORLD DEVELOPMENT

Table 3. Years in which 40 interviewed farmers acquired new capital equipment in Maradi, Nigera
Plough oxen Ox cart Heavy plough Light plough Bicycle Motor cycle
(charrue) (houe)
To 1994 1 8 10 4 2 0
After 1994 18 9 2 34 6 4
Source: Hamadou (2000b).
a
The FCFA was devalued in January 1994, and 1994 was counted in the pre-devaluation epoch, on the supposition
that its effects on prices might not be immediate. (Shaikh & McGahuey, 1996 actually found some pretty immediate
increases in water conservation techniques for cereals in western Niger.) The heavy plough had been promoted, and
credit provided, under a project, which ceased operation in 1987.

overstaffed, inefficient and corrupt) and impel-
ling Senegal to adopt drastic retrenchment
policies (Gaye, 2000).
The important role of livestock, themselves a
form of capital and an insurance substitute, but
also demanding capital for replacement and
maintenance, was evident in all the districts.
But, in lowland Makueni, Kenya, livestock
appeared particularly vulnerable to disease,
and many farmers had experienced substantial
stock losses. Some farmers in the limited areas
with suitable topography had built small dams
to water crossbred cows, to take advantage of
the recently improved milk market when the
Marketing BoardÕs monopoly of urban sales
was abolished, and these also invested in
spraying equipment and veterinary supplies.
For most farmers, dairying was too hazardous
(Fall, 2000; Mbogoh, 2000).
(c) Exports versus local urban markets
Exports overseas have become less important
in the 1990s compared to 1960–70 (Maradi
farmers still export, but to towns just over the
border in Nigeria). Nigerian groundnut exports
fell drastically due to rosette disease in the early
1970s. The crop is now reviving, with the aid of
new resistant varieties, there and in Maradi, but
sales are oriented to the towns (Ariyo et al.,
2001; Hamadou, 2000a). Even in Diourbel,
Senegal, where the Government has continued
to view groundnuts as an export crop, many
farmers grow it, but do not sell (Faye & Fall,
2000), and they rank its uses as food, fodder
and then cash (Faye, 2002). In Makueni, cotton
was briefly promoted as a cash crop under an
aid program 1978–87 but collapsed due to late
payments by the responsible parastatal (Tiffen
et al., 1994). Some farmers would be glad if a
ginnery could be re-established locally (Gichuki
et al., 2000).
(d) Changes in the land market and in income
distribution
Land scarcity has developed everywhere, due
to a slow increase in rural population and im-
proved market access, which has increased the
value of output, and encouraged farm expan-
sion when land was available. As a result, land
has become a marketable commodity, in all
four of the areas examined. This has been least
the case in the Diourbel Region, partly because
of the 1964 Senegal statute ascribing ownership
to the state and user rights to the user, but also
because state policies have in various ways de-
creased the profitability of groundnut and
millet sales, leading to substantial emigration
out of farming 23 (Faye et al., 2001). There has
not been the same competition to obtain land
for young families, or to retain land, as in the
other three districts.
In the other three areas the land market is
active. Intensification increasingly needs cash to
buy labor, inputs, equipment and livestock, in
addition to, or in substitution for, labor for a
delayed reward. Not all farmers have been able
to find the resources and to manage the asso-
ciated risks. Even in Maradi, where intensifi-
cation is at an early stage, some poor farmers
have had to sell fields to the more successful
(Boubacar, 2000). In the Kano area, some
farmers have been less able to cope than others
with the rapid changes in matters such as fer-
tilizer prices, product prices, etc., and Meagher
and Mustapha (1997) noted increased income
differentiation in one village, between 1989–90
and 1992–93, between small farmers who had
to buy in food at rising prices, and larger
farmers, who were better able to play the
markets. Farmers, or their heirs, who inherit
smaller farms, if unable to find the necessary
capital to intensify production from a small
land holding, are increasingly forced to look to
other occupations for a living, and some prefer,
 TRANSITION IN SUB-SAHARAN AFRICA
or are forced, to sell up. Near Sokoto and
Gusau, in northwest Nigeria, some farms have
diminished in size to a garden plot, and the
main income is earned by agricultural wage
labor or other labor. These areas also have
many intermediate viable small farms, and
some large capitalized farms. In the rapidly
changing job and land market of Nigeria, af-
fected by petroleum booms and busts, inflation,
import bans and relaxations, voluntary buyers
and sellers of land may come to regret their
decisions, but the worst sufferers are often those
whose land has been forcibly taken over by the
state (Iliya & Swindell, 1997), often without
compensation, to accommodate urban expan-
sion.
(e) The labor market, specialization and
diversification of incomes
In Nigeria, census data showed average
family size of five in Jigawa, and 5.4 in Kano,
with the (male) household head said to consti-
tute 70–80% of its economically active work
force. 24 In Makueni, the resident family was
5.8, including two adult workers and children
in school. These figures do not suggest an
abundance of labor on family farms. In all four
countries the farm household was also finan-
cially and emotionally linked to adult children
and/or spouses who worked elsewhere either as
commuters, or for part of the year, or as per-
manent urban residents. Family labor has to be
carefully allocated across occupations, and de-
cisions have to be made as to whether to utilize
child labor, or forgo it in favor of sending
children to school.
The model in Figure 3 was built on the
premise that specialization brings benefits. It is
apparent that farmer households in semi-arid
areas do not specialize in the sense of only
growing crops or only rearing livestock, nor in
the sense of relying only on farm income. The
separation of farming and herding made sense
in semi-arid areas in phase A with low popu-
lation density and limited access to markets for
farm products, but in phase B, in semi-arid
areas, the integration of crops and livestock by
smallholders becomes advantageous to provide
manure and draught power, the more so as
their main crops also provide animal feed. 25 In
Maradi, many settled Hausa farmers were ac-
quiring cattle, and many Fulani herders were
cultivating crops around a camp that was be-
coming permanent (CARE International au
Niger & Universite dÕArizona, 1997; Morti-
1359
more et al., 2001). Studies of Fulani livestock
raisers in reserves in northern Nigeria show
both they and the farmers outside the reserve
now combine livestock and cropping, in a va-
riety of ways (Baba & Magaji, 1998; Hassan,
Hoffmann, & Steinbach, 1998). If it is not a
contradiction in terms, semi-arid farmers have
become specialists in mixed farming and crop-
livestock integration.
In semi-arid areas with variable rainfall, in-
come diversification, in the sense of combining
farming and nonfarm activities within one
household, has advantages, particularly by
spreading risk. A Makueni farmer said ‘‘When
there is a drought, I lose as a farmer, but as a
teacher, I never experience that drought. I still
get my salary and can survive better than if I
was only a farmer’’ (Nzioka, 2000, p. 12). Ellis
(2000, p. 5) points out that this cuts across the
view that specialization and market exchanges
are the way to higher incomes and improved
welfare, and it might therefore be thought to be
in contradiction with the economic logic behind
Figure 3. We must distinguish, however, be-
tween household and individual specialization.
In the Kenyan case quoted above, the husband
concentrated on teaching, and his wife on
farming, and they both benefited from sharing
resources. 26 In other cases, a second activity
may be carried out in a different season––a man
farms in the wet season, and trades in the dry,
so he can give full attention to each in turn.
Indeed, by turning over his capital several times
in the dry season, he gains additional resources
to buy inputs for the farming season. 27 It is
necessary again to distinguish between those
who sell small amounts out of necessity, and
those who buy in from others and trade on a
considerable scale. The latter are generally in
the upper echelons of small farmers. CARE
International au Niger and Universite dÕAri-
zona (1997) found grain sold by even in the
most vulnerable households (56% of their Ma-
radi sample) 28 who had then to finance by
some means or another grain purchases later. A
few people in this group and the middle group
(27%) were able to buy on a larger scale, as an
agent of a bigger trader who advanced them
money. Wholesaler traders who bought and
resold hundreds of sacks of grain on their own
account were almost all in the ‘‘riches paysans’’
category (17%).
It has been observed that diversification does
not necessarily level rural incomes. There is
often a U-shaped relationship in which the
proportion of nonfarm in total farm household
1360 WORLD DEVELOPMENT
income is high for both rich and poor, and
lower in the middle groups (Barrett, Bezuneh,
Clay, & Reardon, 2000; Bryceson, 2002; Iliya &
Swindell, 1997; Reardon, Taylor, Stamoulis,
Lanjouw, & Balisacan, 2000; Toulmin et al.,
2000). When diversification is driven by des-
peration, families try to cobble together a living
by farming, working for others, collecting,
hawking, handicrafts etc., and each activity
may impede good achievement in the others.
Ellis (2000) refers to the key difference between
individual diversification and low levels of skill,
associated with poverty, and household diver-
sification, in which individuals specialize, which
is associated with the better off. In Nigeria, the
most successful farmers were those who had
within their household someone in well-paid or
middling formal sector jobs, or in trading and
contracting, which provided a cash flow to fi-
nance farm inputs. These households contained
individual specialists, backed by accumulated
capital or skills, while in poorer households
people scramble for casual labor alongside their
own farming (Iliya & Swindell, 1997). This was
also the case in western Kenya (Hamilton,
2003). In C^ ote dÕIvoire a middle group of spe-
cialist farmers could also be distinguished
(Barrett, Bezuneh, & Aboud, 2001).
In Senegal, Table 3 and Figure 5 show that
the poorer households in Group 1 scraped a
living from a diversity of local occupations. The
wealthier households in Group 3, who were
investing most in livestock, tended to have most
of their nonfarm income in the form of remit-
tances from urban-based children. Those in
residence concentrated on farming, but in all
households, around a fifth of the family work
force was absent in the farming season (Table
3). Farmers in Diourbel calculate the oppor-
tunity cost of using their sonsÕ labor on the
farm or encouraging them to migrate to urban
areas or, if they can manage the higher cost,
overseas. Some of the ceremonial expenses
listed in Table 4 support the Mouride leader-
ship in construction and merchant enterprises
in Dakar, Touba and overseas, where petty
trading, transporting, laboring, etc. provide
income to adherents despite lack of formal
qualifications (Babou, 2002; Wilson Fall, 2000).
A remarkable difference between Kenya and
the two francophone countries was in invest-
ment in education. Makueni parents invested
heavily in educating all their children, boys and
girls, motivated in part by the hope of quali-
fying one or more to compete successfully for a
skilled or professional occupation (Nzioka,
2000). They were well aware that their children
would need nonfarm incomes and, despite high
costs, sent all boys and girls to primary school,
most to nursery school, and 20–30% to sec-
ondary school (with an anxious eye on the job
market, to see if the latter was cost-effective). 29
Education was regarded as a priority expendi-
ture; and farm investments got what was left
over (Nzioka, 2000; Gichuki et al., 2000). On
average the household included two nonresi-
dents, and while 23% of these worked on family
farms elsewhere, 30 26% were in urban em-
ployment in the private sector, and 10% in ur-
ban-based government employment. Due to the
lack of local towns, rural businesses and gov-
ernment services provided comparatively few
jobs (Nzioka, 2000). Educated children in jobs
were expected to, and did, assist their parents
with emergency expenditures, and, importantly,
with financial resources for investments to im-
prove the farm (Nelson, 2000; Nzioka, 2000).
This is household diversification, with individ-
ual specialization. A similar value was placed
on education in western Kenya (Francis, 2000;
Hamilton, 2003).
In Diourbel and Maradi the nonfarm occu-
pations open to the illiterate young were mainly
unskilled labor and petty trading. Rural par-
ents did not see their children as likely to suc-
ceed in academically-oriented schooling, which
starts from day one in French. Hence, migrat-
ing rural children can contribute less to the
productivity of the urban work force in Senegal
and Niger than in Kenya. With unskilled jobs,
their small remittances assist subsistence, not
farm investment (David & Niang, 1995; Diarra
Doka, 2001; Wilson Fall, 2000). Nigeria is an
intermediate position, with some families able
to access jobs through education, and others
investing in traditional clientage relationships
to access less skilled work or to get into trade.
In Nigeria, as in Kenya, nonfarm earnings fi-
nanced farm investments (Iliya & Swindell,
1997).
The market for farm labor has also become
important, as efficient farmers need to buy in
labor at crucial times or for special tasks. In
Makueni we found wages varied according to
demand, level of skill and value of marginal
output (Nelson, 2000). Similar wage variation
was noted in Nigeria, where the agricultural
labor market centers on towns, where it is
easiest for seekers of employment and seekers
of labor to meet. Farmers with very small farms
could intercalate work on these with work for
others, and because of their experience, were
 TRANSITION IN SUB-SAHARAN AFRICA
preferred to cheaper student labor (Swindell
et al., 1999).
5. CONCLUSIONS AND POLICY
IMPLICATIONS
We need to understand the interrelationship
of the agricultural and nonagricultural sectors
at different points in time as population grows,
and redistributes itself between urban and
rural. When the current round of censuses
has been analyzed, we shall probably find more
evidence of rapid change, and governments
need to be ready with appropriate policies. The
growth of the local urban market in SSA has
been demonstrated. Policies which increase the
purchasing power of local urban communities
are becoming more relevant than export-
oriented ones. The recent drops in livestock
turnover in northern Nigeria due to falling
urban incomes (Ariyo et al., 2001) affect not
merely farmersÕ incomes, but also their ability
to insure, and to maintain soil fertility.
Farmers have responded to increased de-
mand by a multitude of small investments
linked to changes in technologies and in prod-
ucts, but they have been hindered by state
control of markets (as in Senegal) or rapidly
changing policies, exchange rates and inflation
(Nigeria, and to a lesser extent, Kenya). Agri-
cultural output can now only be increased by
very efficient combinations of land, labor and
capital. Those farmers who cannot invest are
increasingly being forced out of farming. Some
of the children of even the prosperous farmer
will need, or be attracted to, other livelihoods.
The rural poorest need a vibrant nonfarm sec-
tor.
The appropriate areas for government and
aid-assisted investment need to be identified
bearing in mind both the different types of in-
vestment most required by farm and nonfarm
sectors, and the changing urban–rural balance.
For some countries, such as Niger, the em-
phasis must still be the enabling environment to
allow small farmers to build up their farms, and
become buyers of urban services and manu-
factures. Other countries now need to give
more attention to attracting collective resources
to develop the urban sector, while at the same
time protecting farmersÕ capacity to invest to
meet urban demand. Priorities for constrained
government and aid budgets must be the ser-
vices and infrastructures which enable citizens
to make full use of their own talents, ingenuity
1361
and resources (Drylands Research, 2001).
Communications infrastructure, (roads and
telephones), primary education, law and order,
and good management of the currency are
needed by all sectors, and by farmers as well as
industrialists. The interaction between the two
sectors depends on good communications.
Roads usable by lorries remain vitally impor-
tant both to farmers (Omamo, 1998) and urban
consumers. Education benefits both, if its con-
tent and language are appropriate. Farmers in
Kenya find education, which starts in local
languages, worthwhile for their children, and
invest their own money in building schools and
paying fees (Gichuki et al., 2000; Nzioka,
2000). Swahili, like Hausa, gives access to
written information in newspapers, advertise-
ments, etc. Their adult children and spouses
had a great variety of occupations. Farmers in
the Sahel are not convinced of the benefit of
free education starting in French, but illiter-
acy cuts them off from written information,
(whether on a fertilizer label or in the accounts
of their groupement d’inter^et economique), and
from the better-paying nonfarm jobs.
While farmers can usually find family re-
sources for a profitable investment, some ser-
vices and investments can better be supplied
collectively. Examples are preventative veteri-
nary services, some types of water develop-
ment, and research.
Turning to nonfarm sectors, it is clear that
rural families, and especially the poorer ones,
rely in part on nonfarm income, preferably in a
nearby town or village, if not, in a more distant
city. 31 Premature, state-led industrialization
has a poor record. By 1986 most of the state-
promoted industries in Maradi town, set up in
the late colonial period, or during the uranium
boom had closed, and most employment was in
the informal sector (Gregoire, 1990; Tiffen,
2001). In Senegal, the modern industrial sector
was for various reasons uncompetitive, over-
regulated, and suffered from ‘‘a domestic mar-
ket that was highly restricted by low-incomes’’
(Berthelemy, Seck, & Vourch, 1996, p. 100), so
failed to expand. Nigeria has conditions more
conducive to successful industrialization, and
the modern sector expanded in Kano during
the oil boom, with both state and private cap-
ital, but import-reliant industries suffered in the
1980s from shortages of raw materials, lead-
ing to retrenchment of workers, underutiliza-
tion of capacity, and closures. While sectors
dependent on local raw materials did some-
what better (Olukoshi, 1996), modern industrial
1362 WORLD DEVELOPMENT
establishments fell from 327 in 1989 to 243 by
1993 (Tiffen, 2001). Industry was hindered by
rent-seeking, deteriorating services, and fall-
ing demand from consumers hit by inflation
(Olukoshi, 1996). Runaway inflation and over-
valued exchange rates hindered good manage-
ment. 32 In turn, the downward pressures on
wages reduced effective demand for the higher
value agricultural products.
Productive investment in West Africa has
been more lacking in the industrial sector,
which needs collective investment, than in the
farm sector which, has responded with family
investments. Kenya has a relatively successful
modern sector, but concentrated in a few
NOTES
1. It could be argued that the poorest of the poor could
be helped by income redistribution without growth, but
this is politically a more difficult option.
2. They have been more successful in Asia than
elsewhere. Kriekhaus (2002) has argued that public
savings by states with ‘‘high bureaucratic capacity’’
can successfully allocate resources to strategic in-
dustrial sectors. In many cases, however, the need
to manipulate resources to reward political support-
ers has predominated over what capacity there
might be.
3. Alexandratos (1995) gives a fuller review of devel-
opment policy in regard to agriculture than is appropri-
ate here.
4. There are also higher transaction, insurance and
information costs in scattered markets (Myint, 1985)––
and disruptions from war and disorder.
5. World Bank (2000b, p. 68) quotes a discussion
group in Guatemala: ‘‘We think the earth is generous;
but what is the incentive to produce more than the
family needs if there are no access roads to get produce
to the market?’’
6. Malaysia has made a rapid transition in recent
decades (Gemmell, Lloyd, & Mathew, 2000). Japan
is a well-known example of an earlier rapid
transition. Land-locked countries tend to take
longer than the seaboard countries since initially
the most efficient transport for traded goods was
water-borne.
towns. Government investment in providing
infrastructure to more towns––electricity, piped
water, telecommunications, etc.––may be more
effective than direct state investment in manu-
facture.
Governments cannot do everything, particu-
larly if they are to avoid taxing away private
incentive to invest. They cannot rely on aid,
because aid availability per head is going down.
Hence, as urbanization proceeds, both aid
agencies and national policy makers need to
think seriously about appropriate public in-
vestments for particular countries at a parti-
cular time, and to be conscious that times are
changing.
7. The local black-smithing industry seems to have
responded to growing demand for more efficient metal
tools earlier in West than in East Africa. In Nigeria it is
beginning to be replaced by factory-made equipment
and parts supplied by state organizations (Iliya &
Swindell, 1997).
8. Transport costs consist of fixed and variable costs.
The former are relatively small for human and even
animal transport, but high for lorries. Studies in Malawi
and Benin found traders used logical combinations of
transport, using lorries for distances over 12 and 4 km,
respectively. Motorized transport costs were $0.63 and
$0.28 per ton/km compared with nonmotorized costs
averaging $1.20 and $1.78. Trading costs were also high
in both countries because in the absence of telephones,
traders had to travel personally to obtain information,
place orders, etc. (Fafchamps & Gabre-Madhin, 2001).
High trading costs apply to what farmers need to buy as
well as what they want to sell. Omamo (1998) has shown
how walking more than 8 km to purchase maize or sell
cotton in a market has high opportunity costs, and leads
to a preference prefer to earn cash by migratory labor.
Poor roads reduce farm incomes, ability to supply food
to the towns, and ability to buy goods and services from
urban areas.
9. Exceptions will be industries processing bulky or
perishable raw materials that need to be located near the
source of supply.
10. IBRD (1981) Impact evaluation of the Muda
scheme, consulted by the author for a report to the
then Overseas Development Administration on Improv-
 TRANSITION IN SUB-SAHARAN AFRICA
ing the Socio-economic and institutional content of
irrigation feasibility studies (Research scheme R4006,
ESR. 326/307/01, June 30, 1986).
11. Roads and railways, but also telephones and post
offices.
12. Most francophone countries had their first circa
1977 and another circa 1988. Some have held or planned
a third, at some point 1996–2003 (United Nations
Statistics Division, Population and Housing Census,
http://unstats.un.org/unsd/demographic/census/cendate,
accessed 20/09/02).
13. FAOSTAT gives a more likely 114 million for
2000. The 1963 census, used as the base for the BankÕs
calculations, is widely known to have been manipulated
upward for political reasons.
14. The total agricultural labor force is often estimated
by reference to the population in a given age range, such
as 15–64, from which those known to be in occupations
in the formal sector are deducted. The Bank in its
Technical Notes to employment tables noted some but
not all of the difficulties in estimating unreported
economic activity. It ceased to provide agricultural
labor statistics after 1990, but FAO continues to
estimate them in FAOSTAT.
15. Local food production must have been substantial,
as their food imports in the same year were only $6 and
$7 per capita, respectively.
16. Import data are fairly reliable for countries with
ports, but not all provide it.
17. There is no need to produce more food than can
locally needed, since crops low in value in relation to
bulk can only be exported when transport links are
highly efficient––not usually the case in Africa.
18. The sample surveys were not intended to collect
data at a level to enable significance testing, but rather to
check and update the information available from the
district and the literature.
19. It is not just surplus food that is sold, as some
households sell off grain after harvest or at other times to
meet urgent needs, buying back, often at a disadvantage
price-wise, later on.
20. WoteÕs improved administrative status led to a
telephone service and some electricity by 2000.
1363
21. Statistics were only available from 1992, when the
District was formed. Agricultural officers were not able
to get out much to estimate production due to budgetary
constraints.
22. Fertilizer was still available at official prices from
state organizations in the 1990s, but tended to be
monopolized by big farmers or those with the right
connections, who might then sell on at higher prices.
Subsidies benefited mainly the big farmers (Iliya &
Swindell, 1997; Swindell et al., 1999).
23. Our Senegalese colleagues were particularly keen to
test this finding, which had depended mainly on extra-
polation from census data, and some reports––Tiffen in
Barry et al. (2000). Villagers confirmed it.
24. This takes little account of the income-earning
activities of wives, but in Hausa areas these generally
prefer to pursue their own crafts rather than to work
unpaid on the husbandÕs farm.
25. As population density rises, and more pasture land
is converted to crop production, farmers need manure to
replace fallows, and herders find the advantage of
combining farming and herding from a settled base, a
change analyzed by Binswanger, McIntire, and Udry
(1989).
26. There are also cases in Makueni where the husband
works in the farm and the wife is a nurse or teacher.
27. As reported to the author by a farmer in Gombe,
Nigeria, in the 1960s.
28. CARE interviewed men and women in more than
400 households throughout Maradi District, in 1996.
29. By 1998, there were slightly more girls in sec-
ondary. Parents having difficulty with the fees might
think sons had a better chance of earning without
formal qualifications (Nzioka, 2000; Gichuki et al.,
2000).
30. Sons are often given a part of the farm to use on
their marriage (but the parent retains the title deed).
31. Long-distance migration seems more disruptive to
family structure, judging by Francis (2000).
32. Olukoshi (1996) regarded the stabilization of the
macro-policy environment as a necessary condition for
the reversal of de-industrialization.
1364 WORLD DEVELOPMENT
REFERENCES
Alexandratos, N. (1995). World agriculture towards
2010: An FAO study. Chichester: FAO and John
Wiley & Sons.
Anderson, D. (1990). Investment and economic growth.
World Development, 18(8), 1057–1079.
Ariyo, J. A., Voh, J. P., & Ahmed, B. (2001). Long-term
change in food provisioning and marketing in the
Kano Region. Drylands Research Working Paper
34. Crewkerne, UK: Drylands Research.
Baba, K. M., & Magaji, M. D. (1998). Fadama crop
residue production and utilisation in north-western
Nigeria. In I. Hoffmann (Ed.), Prospects of pastoral-
ism in West Africa (pp. 247–262). Giessen, Germany:
Tropeninstitut.
Babou, C. A. (2002). Brotherhood solidarity, education
and migration: The role of the Dahiras among the
Murid muslim community of New York. African
Affairs, 101(403), 151–170.
Balcet, J.-C. (1997) Retrospective assessment of farm
technology adoption in northern Nigeria (1982–
1997): emerging lessons for technology development
and the rural development agenda. Unpublished
report. Washington, DC: World Bank.
Barrett, C. B., Bezuneh, M., & Aboud, A. (2001).
Income diversification, poverty traps and policy
shocks in C^ ote dÕIvoire and Kenya. Food Policy,
26, 367–384.
Barrett, C. B., Bezuneh, M., Clay, D. C., & Reardon, T.
(2000). Effects of non-heterogeneous constraints,
incentives and income diversification strategies in
rural Africa. Mimeo. Department of Agricultural,
Resource and Managerial Economics, Cornell Uni-
versity.
Barrett, C. B., Reardon, T., & Webb, P. (2001).
Nonfarm income diversification and household live-
lihood strategies in rural Africa––concepts, dynamics
and policy implications. Food Policy, 26(4), 315–331.
Barry, A., Ndiaye, S., Ndiaye, F., & Tiffen, M. (2000).
Region de Diourbel: Les aspects demographiques.
Drylands Research Working Paper 13. Crewkerne,
UK: Drylands Research.
Bell, C., Hazell, P. B. R., & Slade, R. (1982). Project
evaluation in regional perspective: A study of an
irrigation project in NW Malaysia. Baltimore: John
Hopkins University.
Berthelemy, J.-C., Seck, A., & Vourch, A. (1996).
Growth in Senegal: a lost opportunity? Paris: OECD.
Binswanger, H., McIntire, J., & Udry, C. (1989).
Production relations in semi-arid African agricul-
ture. In P. Bardhan (Ed.), The economic theory of
agrarian institutions (pp. 122–144). Oxford: Claren-
don Press.
Boserup, E. (1965). The conditions of agricultural growth:
The economics of agricultural change under population
pressure. London: Allen and Unwin.
Boubacar, Y. (2000). E  volution des regimes de propriete
et dÕutilisation des ressources naturelles dans la
region de Maradi. Drylands Research Working
Paper 29. Crewkerne, UK: Drylands Research.
Boubacar, Y., & Ibrahim, B. M. (2002) Rapport de
mission de restitution des travaux de recherche dans
les villages de Jiratawa, Magami, Sharken-Hausa et
Dan-Kullu (Departement de Maradi, Niger). Un-
published report. Crewkerne, UK: Drylands Re-
search.
Bryceson, D. F. (2002). The scramble for Africa:
reorienting livelihoods. World Development, 30(5),
725–739.
CARE International au Niger & Universite dÕArizona
(1997). Evaluation de la securite des conditions de vie
dans le Departement de Maradi. Niamey, Niger:
CARE-International au Niger.
Coulon, C. (1999). The Grande Magal in Touba: a
religious festival of the Mouride Brotherhood of
Senegal. African Affairs, 98(391), 195–210.
David, R., & Niang, O. M. (1995). Diourbel, Senegal. In
R. David (Ed.), Changing places? Women, resource
management and migration in the Sahel (pp. 23–53).
London, UK: SOS Sahel.
Diarra Doka, M. (2001). E  volutions a long terme de
lÕorganisation sociale et economique dans la region
de Maradi. Drylands Research Working Paper 26.
Crewkerne, UK: Drylands Research.
Drylands Research (2001). Livelihood transformations in
semi-arid Africa 1960–2000: Proceedings of a work-
shop arranged by the ODI with Drylands Research
and the ESRC, in the series Ôtransformations in
African agriculture.Õ Drylands Research Working
Paper 40. Crewkerne, UK: Drylands Research.
Ellis, F. (2000). Rural livelihoods and diversity in devel-
oping countries. Oxford: Oxford University Press.
Fafchamps, M., & Gabre-Madhin, E. (2001). Agricul-
tural markers in Benin and Malawi: The operation
and performance of traders. Policy Research Work-
ing Paper 2734. Washington, DC: World Bank
Development Research Group, Rural Development.
Fall, A. (2000). Makueni District profile: livestock
management, 1990–1998. Drylands Research Work-
ing Paper 8. Crewkerne, UK: Drylands Research.
Faye, A. (2002). Comte rendu de la journee de restitu-
tion: Recherches sur les politiques publiques et les
investissements paysans dans la Region de Diourbel;
CNRA de Bambey, 19 fevrier 2002. Unpublished
report. Crewkerne, UK: Drylands Research.
Faye, A., & Fall, A. (2000). Region de Diourbel:
Diversification des revenus et son incidence sur
lÕinvestissement agricole. Drylands Research Work-
ing Paper 22. Crewkerne, UK: Drylands Research.
Faye, A., Fall, A., & Coulibaly, D. (2000). Region de
Diourbel: evolution de la production agricole. Dry-
lands Research Working Paper 16. Crewkerne, UK:
Drylands Research.
Faye, A., Fall, A., Mortimore, M., Tiffen, M., & Nelson,
J. (2001). Region de Diourbel: synthesis. Drylands
Research Working Paper 23e. Crewkerne, UK:
Drylands Research.
Francis, E. (2000). Making a living: Changing livelihoods
in rural Africa. London: Routledge.
 TRANSITION IN SUB-SAHARAN AFRICA
Gaye, M. (2000). Region de Diourbel: politiques natio-
nales affectant lÕinvestissement chez les petits explo-
itants. Drylands Research Working Paper 12.
Crewkerne, UK: Drylands Research.
Gemmell, N., Lloyd, T. A., & Mathew, M. (2000).
Agricultural growth and inter-sectoral linkages in a
developing economy. Journal of Agricultural Eco-
nomics, 51(3), 353–370.
Gerschenkron, A. (1962). Economic backwardness in
historical perspective. Cambridge, MA: Harvard
University Press.
Gichuki, F. N. (2000a). Makueni District profile:
farm development, 1946–1999. Drylands Research
Working Paper 1. Crewkerne, UK: Drylands
Research.
Gichuki, F. N. (2000b). Makueni District profile: soil
management and conservation, 1989–1998. Drylands
Research Working Paper 4. Crewkerne, UK: Dry-
lands Research.
Gichuki, F. N., Mbogoh, S. G., Tiffen, M., & Morti-
more, M. (2000). Makueni District profile: synthesis.
Drylands Research Working Paper 11. Crewkerne,
UK: Drylands Research.
Gregoire, E. (1990). Croissance urbaine et sante a Maradi
(Niger). Bordeaux: Universite de Bordeaux II.
Hamadou, S. (2000a). E  volution a long terme des
productions agricoles, du systeme de commercialisa-
tion et des prix des produits dans la zone de Maradi.
Drylands Research Working Paper 32. Crewkerne,
UK: Drylands Research.
Hamadou, S. (2000b). Politiques nationales et investiss-
ement dans les petites exploitations a Maradi.
Drylands Research Working Paper 33. Crewkerne,
UK: Drylands Research.
Hamilton, P. (2003). Struggling to survive poverty: A
survey of small farmersÕ coping strategies in rural
Kenya. ENABLE: Newsletter of the Association for
better land husbandry, 16(March), 3–30.
Hassan, W. A., Hoffmann, I., & Steinbach, J. (1998).
Sheep and goat keeping under arable farming in the
Zamfara Reserve. In I. Hoffmann (Ed.), Prospects of
pastoralism in West Africa (pp. 196–210). Giessen,
Germany: Tropeninstitut.
Helleiner, G. K. (1966). Peasant agriculture, government,
and economic growth in Nigeria. Homewood, IL:
Irwin.
Hoffmann, I. (1998). Cattle marketing in Burkina Faso
after the devaluation of the Franc CFA. In I.
Hoffmann (Ed.), Prospects of pastoralism in West
Africa (pp. 127–135). Giessen, Germany: Tropenin-
stitut.
Hopkins, A. G. (1973). An economic history of West
Africa. London: Longman.
Iliya, M. A., & Swindell, K. (1997). Winners and losers:
household fortunes in the urban peripheries of
northern Nigeria. In D. F. Bryceson, & V. Jamal
(Eds.), Farewell to farms: De-agrarianisation and
employment in Africa. Aldershot, UK: Ashgate.
Jahnke, H. E. (1982). Livestock development systems and
livestock production in tropical Africa. Kiel: Wis-
senschaftsverlag Vauk.
1365
Johnston, B. F., & Mellor, J. W. (1961). The role of
agriculture in economic development. American
Economic Review (September), 566–593.
Kenny, C., & Williams, D. (2001). What do we know
about economic growth? Or, why donÕt we know
very much? World Development, 29(1), 1–22.
Kherallah, M., Delgado, C., Gabre-Madhin, E. Z.,
Minot, N., & Johnson, M. (2000). The road half
traveled: Agricultural market reform in sub-Saharan
Africa. Washington, DC: International Food Policy
Research Institute.
Kriekhaus, J. (2002). Reconceptualizing the develop-
mental state: Public savings and economic growth.
World Development, 30(10), 1697–1712.
Landes, D. (1998). The wealth and poverty of nations.
London: Little, Brown & Co.
Mahamane, A. (2001). Usages des terres et evolutions
vegetales dans le departement de Maradi. Drylands
Research Working Paper 27. Crewkerne, UK: Dry-
lands Research.
Mbogoh, S. G. (2000). Makueni District profile: crop
production and marketing, 1988–1999. Drylands
Research Working Paper 7. Crewkerne, UK: Dry-
lands Research.
Meagher, K. (1997). Current trends in cross-border
grain trade between Nigeria and Niger. Paris:
IRAM.
Meagher, K., & Mustapha, A. R. (1997). Not by
farming alone: the role of non-farm incomes in rural
Hausaland. In D. Bryceson, & V. Jamal (Eds.),
Farewell to farms: De-agrarianisation and employ-
ment in Africa. Aldershot, UK: Avebury Ashgate
Publishing Ltd.
Mortimore, M. (1993). The intensification of peri-urban
agriculture: the Kano Close-Settled Zone, 1964–86.
In B. L. Turner, R. W. Kates, & H. L. Hyden (Eds.),
Population growth and agricultural change in Africa.
Gainesville: University Press of Florida.
Mortimore, M., & Adams, W. (1999). Working the
Sahel: environment and society in northern Nigeria.
London: Routledge.
Mortimore, M., Tiffen, M., Boubacar, Y., & Nelson, J.
(2001). Department of Maradi: synthesis. Drylands
Research Working Paper 39e. Crewkerne, UK:
Drylands Research.
Mustapha, A. R., & Meagher, K. (2000). Agrarian
production, public policy and the State in Kano
Region, 1900–2000. Drylands Research Working
Paper 35. Crewkerne, UK: Drylands Research.
Myint, H. (1967). The economics of the developing
countries. London: Hutchinson University Library.
Myint, H. (1985). Organisational dualism and economic
development. Asian Development Review, 3(1), 25–
42.
Nelson, J. (2000). Makueni District profile: income
diversification and farm investment, 1989–1999.
Drylands Research Working Paper 10. Crewkerne,
UK: Drylands Research.
Nzioka, C. (2000). Makueni District profile: human
resource management, 1989–1998. Drylands Re-
search Working Paper 9. Crewkerne, UK: Drylands
Research.
1366 WORLD DEVELOPMENT
OÕBrien, D. B. C. (1975). Saints and politicians: Essays in
the organisation of a Senegalese peasant society.
Cambridge, UK: Cambridge University Press.
Okike, I., Jabbar, M. A., Manyong, V., Smith, J. W.,
Alinwumi, J. A., & Ehui, S. K. (2001). Agricultural
intensification and efficiency in the West African
savannahs: evidence from northern Nigeria Socio.
Economics and Policy Research Working Paper 33.
Nairobi, Kenya: ILRI.
Olukoshi, A. (1996). Economic crisis, structural adjust-
ment and the coping strategies of manufacturers in
Kano, Nigeria Discussion Paper 77. Geneva: UN-
RISD.
Omamo, W. S. (1998). Transport costs and smallholder
cropping choices: an application to Siaya District,
Kenya. American Journal of Agricultural Economics,
80, 116–123.
Orr, A., & Mwale, B. (2001). Adapting to adjustment:
smallholder livelihood strategies in southern Malawi.
World Development, 29(8), 1325–1343.
Pingali, P., Bigot, Y., & Binswanger, H. (1987). Agri-
cultural mechanization and the evolution of farming
systems in sub-Saharan Africa. Baltimore: Johns
Hopkins University Press.
Prabowo, D., & McConnell, D. J. (1993). Changes and
development in Solo Valley farming systems. Indone-
sia, Rome: FAO.
Reardon, T. (1997). Using evidence of household income
diversification to inform study of the rural nonfarm
labor market in Africa. World Development, 25(5),
735–747.
Reardon, T., Taylor, J. E., Stamoulis, K., Lanjouw, P.,
& Balisacan, A. (2000). Effects of nonfarm employ-
ment on rural income inequality in developing
countries: an investment perspective. Journal of
Agricultural Economics, 51(2), 266–288.
Republic of Niger (1992). Recensement de la population
de 1988: Analyse des donnees definitives.
Romer, P. M. (1989). What determines the rate of
growth and technological change? Country Econom-
ics Department WPS. Washington, DC: World Bank.
Rostow, W. W. (1960). The stages of economic growth.
Cambridge, UK: Cambridge University Press.
Schoonmaker Freudenberger, M., & Schoonmaker
Freudenberger, K. (1993). Fields, fallow and flexi-
bility: natural resource management in Ndam Mor
Fademba, Senegal. Drylands Paper 5. London:
IIED.
Shaikh, A., & McGahuey, M. (1996). Niger case study:
capitalizing on change. In D. C. Clay, T. Reardon, &
A. Shaikh (Eds.), Population, environment and devel-
opment in Africa: Dynamic linkages and their impli-
cations for future research and development
programming (pp. 93–118). Michigan State Univer-
sity: Institute of International Agriculture.
Simon, J. (1977). The economics of population growth.
Princeton: Princeton University Press.
Singh, B. B., Larbi, A., Tabo, R., & Dixon, A. G. O.
(2001). Improvement for crop-livestock systems de-
velopment in West Africa. ILRI-IITA, Ibadan,
Nigeria.
Smith, A. (1776). The wealth of nations. London:
EverymanÕs Library, 1991, Random House (UK)
Ltd. (Reprinted).
Smith, J., Barau, A. D., Goldman, A., & Mareck, J. H.
(1994). The role of technology in agricultural inten-
sification: the evolution of production systems in the
northern guinea savanna of Nigeria. Economic
development and cultural change, 42(3), 537–554.
Swindell, K., Iliya, M. A., & Mamman, A. B. (1999).
Making a profit, making a living: commercial farm-
ing and urban hinterlands in north-west Nigeria.
Africa, 69(3), 386–403.
Tiffen, M. (2001). Profile of demographic change in the
Kano–Maradi Region, 1960–2000. Drylands Re-
search Working Paper 24. Crewkerne, UK: Drylands
Research.
Tiffen, M., & Mortimore, M. (1994). Malthus contro-
verted: the role of capital and technology in growth
and environment recovery in Kenya. World Devel-
opment, 22(7), 997–1010.
Tiffen, M., & Mortimore, M. (2002). Questioning
desertification in dryland sub-Saharan Africa. Nat-
ural Resources Forum, 26, 218–233.
Tiffen, M., Mortimore, M., & Gichuki, F. (1994). More
people less erosion: environmental recovery in Kenya.
Chichester, UK: John Wiley & Sons.
Tomich, T. P., Kilby, P., & Johnston, B. F. (1995).
Transforming agrarian economies: Opportunities
seized, opportunities missed. Ithaca: Cornell Univer-
sity Press.
Toulmin, C., Leonard, R., Brock, K., Coulibaly, N.,
Carwell, G., & Dea, D. (2000). Diversification of
livelihoods: evidence from Ethiopia and Mali. IDS
Research Report, No. 47. Brighton, UK: Institute of
Development Studies.
Wiggins, S. (1995). Change in African farming systems
between the 1970s and the 1990s. Journal of Inter-
national Development, 7(6), 807–848.
Wiggins, S. (2000). Interpretating changes from the
1970s to the 1990s in African agriculture through
village studies. World Development, 28(4), 631–662.
Wilson Fall, W. (2000). Region de Diourbel: the family,
local institutions and education. Drylands Research
Working Paper 20. Crewkerne, UK: Drylands Re-
search.
World Bank (1983). World Development Report 1983.
Washington, DC: World Bank.
World Bank (2000a). Can Africa claim the 21st Century?
Washington, DC: World Bank.
World Bank (2000b). World Development Report 2000/
2001: Attacking poverty. New York: Oxford Uni-
versity Press.
World Bank (2000c). World Bank Africa Database 2000.
CD-Rom. Washington, DC: World Bank.
World Bank (2002). World Bank Africa Database 2002.
CD-Rom. Washington, DC: World Bank.
World Bank, O.E.D (1995). Nigeria, Impact Evaluation
Report, Kano Agricultural Development Project,
Sokoto Agricultural Development Project. No.
14767-UNI. Washington, DC: World Bank.